Rank-Minimization with a two-step analysis should replace randomization in clinical trials

摘要

Several reviews over the past 15 years (Treasure and MacRae [1], Scott et al. [2], Blair [3] andTaves [4]) have concluded that minimization is superior to randomization. However, acceptance of minimization has been mixed because of its vulnerability to selection bias [5] and the uncertainty about how to analyze the data. It is thus not surprising that although the number of articles that report using minimization has tripled over the past decade, it still represents less than 2% of all clinical trials [4]. Rank-Minimization [6] together with new ways to control selection bias and new procedures for statistical analysis make it appropriate to consider replacing the use of randomization in all clinical trials.